Slack as an ephemeral communication medium

Continuing the discussion from Bulk update of pedestrian crossings by LocalMapper:

This is true. Currently, such a scenario is only a hypothetical, but we don’t only have to be vigilant for something as formal as a change in the terms of use. Recently, Salesforce overhauled Slack’s user interface, and some people are struggling to adapt. (The default hot aubergine color scheme is not really my cup of tea.) I think we’ll eventually get used to Slack’s new user interface, but it is a wake-up call.

In a recent Slack thread, a concern was raised about the portability of all the discussions that have taken place on Slack. Fortunately, the same plan that gives OSMUS full searchability also allows the workspace’s administrators to export a full dump of all the message history. However, there’s no plan to publish this message history as a public archive. Here’s how I explained that stance (edited for intelligibility):

This is partly a privacy/safety issue: many people have joined this workspace over the years with an expectation that their writings would not be exposed publicly for Google and now ChatGPT to get ahold of. It would be problematic to backtrack on that understanding wholesale. At least we’d have to ask for everyone’s permission, which would be a logistical challenge. If worst comes to worst and we have to move off Slack, it might be feasible to archive the chat history and make it accessible to those who have joined this workspace but not the whole wide world, either on a successor platform or a custom website thingie.

Regarding the incongruence of an open data community gathering on a closed platform:

This workspace has an open door but still has walls. I think most people would be OK with most of their writing being available to the rest of the OSM community, but, like, do you remember what questions you asked and which jokes you tried to land when you first joined? I sure don’t. :sweat_smile:

But it doesn’t have to be this way forever:

Over time, I think it would be healthy to summarize some of the more substantive discussions (and maybe retell some of the more successful jokes) on the forum or wiki, so that we can all have the same attitude as some others have about this workspace’s ephermerality.

And, presciently:

“It happened on Slack/IRC” is not always a very convincing argument when defending a tagging decision with someone who isn’t on Slack/IRC.


And wearing my OSM US board member hat, I would oppose publishing such an archive for that basic reason.

Precisely. And the existence of that door, though open to anyone who chooses to walk through it, still acts as a filter for who chooses to step inside. Each communication space on the project – the free-as-in-speech ones, the free-as-in-beer ones, and the closed ones, all have a unique feel and culture based on the people, technology and circumstances of the space. A project like ours flourishes more when contributors have a variety of participation options that suit their philosophy and ideology.

1 Like

On that, does anybody know if discussions here are “visible” to Google / ChatGPT?

With my DWG hat on, it would be really handy if all discussions were available / visible in one place, so we don’t ask the question of “Why did you do this without talking about it first?”, but then get “Oh, but we talked about on Slack / Telegram / Mastodon / Discord / Facebook / etc, & all the members of our private group there thought it was a great idea!” :roll_eyes:

Slack, Telegram, Discord, IRC, and other chat methods are not a good way to do consultation with the community. By design, messages are somewhat ephemeral in nature, even if logged somewhere. Most users on the platform will not keep up with all messages when they were away, and expecting them to do so is misusing it.

A standard use of chat would be to ask a question, interact with others about the question until I have an answer, then disengage until I next chat. At this point, all I know is that I have an answer, not that the community has had a chance to comment. A community member who was away for a day can’t be expected to read all messages, particularly on a high-traffic server.

Threads, like Slack and Discord have, improve this slightly, but don’t change the underlying real-time nature of the platform.


Google does index this forum, though its coverage of the mailing lists has dramatically declined over the years, so I’d expect its coverage of the forum to be similarly scattershot.

Anyhow, this was a rhetorical device; the point is that many people, especially those on the periphery of this project, feel more comfortable in an environment sheltered from the prying eyes of social media and search engines. The culture of civility and empathy there exemplifies the best in OSM. But because of the login wall and fast-paced action, it’s better-suited to sharing advice, links, and funny photos than long-term technical decisionmaking.

If someone did get a big :+1::+1: from the community on Slack, they should be able to produce a link to the discussion. Best if they can provide that URL upfront in their changeset tags. Five DWG members are members of the workspace and can ensure that the mapper didn’t just ask ChatGPT to fabricate a Slack URL for them and that they didn’t post in #all-alone without getting a response. I recall an episode a year or two ago when someone on the talk-us list tried to pass off Slack as having already acquiesced to their import idea; they were quickly called out for it.

At the end of the day, someone making a major edit is responsible for doing due diligence and hearing out a range of perspectives in whichever venue can facilitate that dialogue. On a practical level, they would need to somehow involve the OSMUS Slack workspace, even if it’s just cross-posting a link to a forum or wiki discussion, because Slack dwarfs any of the other relevant communication channels (edited with stats from the summer doldrums):

But I want to emphasize that, even though the atmosphere is very different on Slack than the mailing lists or forums have historically been known for, we’re all part of the same project and our shared goals overlap much more than we give ourselves credit for. The sky didn’t fall when we recently started allowing import discussions to take place on the forum instead of the mailing lists. Nor did it fall several years earlier, when the majority of active talk-us participants invited themselves into what had been the OSMUS board’s private Slack workspace for day-to-day work, simply because it met the pent-up demand for real-time chat.

This forum did exist in some form back when the community reached a consensus with their feet that Slack was the better alternative. Times are changing, but I think we can work towards a more open recordkeeping practice without rushing to delegitimize the consensuses that we’ve achieved on purely procedural grounds.

& I’m one of them, but only ever go on there if I need to ask an US specific question.

Which was never my intention!

1 Like

This is part of the story, and I’ve made the same point myself, but I think it overlooks the fact that the Slack workspace has a lot of people from every part of the country who are actively trying to solve problems for each other.

Recently, I was commiserating with another mapper who, like me, frequents the OpenStreetMap Vietnam group on Facebook. A Facebook group’s form factor resembles that of a forum much more than a chat room. However, it’s very difficult to get a thoughtful response to a serious question there. The group includes some hard-core mappers, but if I ask a tagging question there, the best case would be a collective shrug; I’d just have to ATYL and hope for the best.

It’s no accident that Slack is marketed as enterprise business software. It may be chat, but people sign up there to be productive members of the OSM community. The best thing we can do is find ways to better integrate what happens there with the rest of the OSM experience. As far as I’m concerned, the login wall can stay up, but I think there are many mappers for whom Slack can be a steppingstone to the forum, wiki, and GitHub repositories.

well that’s a legit point. But can OSMUS at least ensure that they have an export of the data? Just in case the rug gets pulled and it’s too late (“oops no-one did an export before they turned off the export functionality, now 10+ years of OSM history are gone”). We’ve seen with Twitter and Maxar that things can change.


I haven’t been involved as much, but I will just add my opinion for a number of reasons it’s better to have the Discourse forums as an officially sanctioned communication platform. I think the reasons have been stated many times already so I won’t get into them too much (open platform is better than a closed source platform that we are “allowed to use” for free) I don’t visit the Slack very much anymore and in the few times I have posted things in my local Washington channel, I didn’t get much engagement.

Whether or not there is an export of the Slack, it’s just not clear what the benefit is to have it. I don’t think there is a reasonable expectation of privacy, but perhaps some people used private messages and would have the expectation those would not somehow become public. Exporting the data and keeping it around risks leaking that.

1 Like

Speaking only for myself and not the rest of the board: no, we are not doing that. The fact that chat is ephemeral is a feature, not a bug.

Slack is a real-time chat platform more akin to IRC, not a deep archive of the project. If it “goes away”, we would of course have to find a new service to use for real-time chat (of which there are many options), and I’m sure this would be very annoying. But even in that case, we would not import Slack history into some hypothetical successor.

While folks have found it useful to be able to deep link into Slack, nobody should assume that these are forever links. If you want permanent linkable content, use this Discourse server or the wiki.

1 Like

To be clear, I agree with this stance, and I have actively advised mappers to create a public paper trail when it comes to anything consequential. However, for the time being, I also defend mappers who cite Slack conversations as justification for their edits, because they are literally following the guidelines. If we reach a consensus that the guidelines are outdated, then we change the guidelines for decisions going forward, and recreate that paper trail to the extent possible. Unfortunately, every communication medium has an uneven track record when it comes to establishing consensus, even without the element of ephemerality.


For those who missed it. OSMUS might have to start paying $6k→$80k (?) p.a. to keep the history of the slack server. AIUI, the free version is limited to 90 days of history.

Looks like there might be an emphermal version of the OSMUS slack in any case!


You’re making it look like I said that in a non-sarcastic way :slight_smile: and have removed “, but “indefinitely” is a very long time. In the short life of the internet, this has happened to platform after platform” from what I said to do that!

1 Like

Just to set the record straight, Salesforce is quoting us $13 per year, per active user beyond some threshold (so the first X are free), measured every 28 days. At last check, Slack had about 800 active users by this definition (out of the 6200 current users). So the ballpark is closer to the $8K range per year (but with the risk that the cost goes up as the number of users do). We would have to sign up for a year under these terms, which is a massive discount over the $87 per year that they charge commercial customers.

So, if you are an active Slack user, and OpenStreetMap US Member, thank you for your support. Your annual membership fee, starting at $20 for regular members, helps us pay for all sorts of programming, such as the virtual Mapping USA conference which starts tomorrow. If you’re an active Slack user but you aren’t an OSM US member, joining is one way you can help support the organization.

So from my OSM US board member perspective, our options are:

  • Pay the cost of Slack and reduce our programming in some other area
  • Drop down to the free tier and lose a number of Slack features such as history older than 90 days and certain integrations
  • Switch to an alternative but pay someone to set it up, host, and administer it
  • Switch to an alternative but someone in the community volunteers to do the work of investigating alternatives and administering it

My primary concern is making sure that the community we’ve built up has a place to go be productive and collaborate on this free map of the world that we all care about.

Slack provides some pretty neat features that set it apart from other chat services. For example, Discord and IRC are absolutely not viable alternatives for the Slack community, but there are other possibilities.

The key features in my opinion of Slack (not an exhaustive list) are:

  • Users can create arbitrary channels
  • Direct and group private messages
  • Ability to create private channels
  • Deep linking to messages
  • RSS integration (lets us link OSMCha filters to a chat channel, for example)
  • Inline threading
  • Share messages between channels
  • Share thread comments to the main channel
  • Full emoji support (including my personal favorite :popcorn:)
  • Users can add their own arbitrary emoji reactions
  • Share files

Now, we wouldn’t need 100% feature parity, but these are all collaboration features that Slack users have come to expect and should be considered when evaluating alternatives.

1 Like

For what it’s worth, almost all of these are possible on Discord out of the box. The exceptions are RSS feeds and allowing users to create channels, which can be done with bots. I understand that there are plenty of reasons outside of that list that OSM US would be wary to switch to discord, though.

1 Like

For now, the comparison to IRC is inaccurate because we actually have a deep archive of Slack, unlike with IRC. I also don’t think there’s much of a chance that the chat history will ever be transferred to another chat platform due to privacy concerns, as not everyone will make the leap and may not want their writings accessible to essentially a different community under different terms.

I think preserving history in that sense would be foolish anyways: inevitably, the next chat platform will also experience scaling issues or enshittification and we’ll find ourselves as nomads once again. To the extent that the history matters, we could find ways to keep a snapshot of the existing archive accessible to Slack users or some subset thereof, for peace of mind, but otherwise we should make a break with the past and move on.

I appreciate the brave talk about completely walking away from the chat history, but I think setting up a static archive that OSMUS controls would make a migration to a different platform more realistic, because we can then move to Slack’s free tier as a forcing function for treating Slack as an ephemeral medium. It takes off the table the main asset that Salesforce leverages to lock us in and charge us real money. It would buy the community time to do all these aspirational things about documentation at our own pace, since it’s not like we’re all employees who can be assigned that work on a deadline.

Unfortunately, in the last few years, it has become a not uncommon practice to cite a Slack discussion by URL or even just by a vague reference inside of a changeset comment, as justification for a change that, by the mapper’s own admission, was significant enough to require consulting the community.

While we shouldn’t be demanding a permanently publicly inspectable chain of custody regarding every decision about every edit, we unfortunately have seen many cases where someone comes along much later and tries to relitigate a decision on purely procedural grounds. I think these changesets demonstrate the point that we need to reduce our reliance on chatrooms for decisionmaking. Any chatroom should only be a place to brainstorm, and maybe arrive at a consensus if it’s truly not a big deal, but it needs to be documented somewhere more durable.

I tend not to go out on such a limb that I need to cite discussions as “sources” on changesets, so to me it’s usually enough just to explain the objective rationale directly in the changeset comment. But if you feel such an urge to have a consensus back you up, that consensus should be more meaningful than “per Slack” and more human-readable than or If your changes really have merit, you should believe in them without hiding behind an anonymous link.

By way of an update, the guidelines have changed to deemphasize Slack and other chat platforms:


I’m a moderator on the OSM World Discord, and I use it all the time. The Discord threading model is incredibly poor and I would not consider it to be an acceptable replacement. Replies go inline by default and the thread UI is terrible. Slack has a nice clean interface that shows you at a glance the length of thread and who’s on it. The bot situation feels like an unholy mess of hacks on top of hacks, and to top it off, I have to run an app to use Discord whereas Slack I can access via a web browser. Don’t get me wrong, we have a great community there, but it’s not a viable Slack replacement.

On the OSM World Discord server we have apps which post from RSS feeds, e.g. the mastodon social media posts tagged '#OpenStreetMap`, OSM diary entries, or State of the Map news.

As a user of both, I find threads in Slack both a great strength and also a bit of a weakness once all the discussion is in a thread and it’s veered off-topic (which large threads inevitably do).

The Discord client is pretty user-hostile, but in the case of the OSM World server, the channels are bridged to Matrix so you can use one client for IRC, Discord and (if any existed of interest) Matrix.

1 Like